Lightsaber Crystals

This is a pseudo guest post by my son (aka "the Boy"). He and a friend were in the backyard dueling with their "build-your-own" lightsabers from Disney World. In an epic exchange, one of the lightsabers fell apart and a little door popped open revealing three plastic crystals in the flashlight handle. When you pull them out in different combinations, the energy vibration sound changes. We've had these toys for a few years and this is the first we've known of this. There are a few comments on blogs around the web, but not a whole lot of info is out there. "Dad, will you put this on your blog so we can get the word out?" So here we are.

To find the crystals, take off the grip pieces from the flashlight handle, like you're going to change the batteries. Right next to the battery compartment, is the crystal chamber. Here's step-by-step photographs showing the sequence:

The crystals pull out by the top edge where there's a little ridge. They go back in easily the same way. If you have trouble reinserting them, rotate them around and make sure they are in the right slot. It's not too hard to force them in the wrong way making it difficult to pull them back out. Here's a close up of the compartment with and without the crystals:

With three crystals, there are eight possible combinations, but turns out there are only three different sounds. Here's a truth table with links to the sound files.

Extra credit if you draw the logic circuit to implement the truth table.

HO Trains and DCC++: Part 2

Controlling a DCC++ base station using JMRI and WiThrottle on an iPad.

At breakfast this morning the Boy asked, “Dad, can we fix Gordon today?” Gordon, from Thomas and Friends, didn’t work when the trains were rediscovered. So today, we took him apart to diagnose the problem. He simply wouldn’t run. Turns out the fix was easy: the contact leafs that rub against the drive wheels to pick up power needed a little extra bending to make a low resistance connection. Once we did that, we also had to run him out a little - I suppose the lubricant in the gear box needed to be redistributed. That and little machine oil and he was up and going. The Boy wanted to put this on the blog. Here is a picture of Gordon in a state of disassembly.

Last time, I was setting up a DCC++ base station to control our one DCC locomotive. Since then, I installed JMRI on a netbook running Xubuntu to have some better control. JMRI is a Java-based train control system that supports DCC++ for output. My netbook was running XP and wouldn't support a recent JMRI version, so I installed Xubuntu as a dual boot. JMRI installation on Xubuntu was pretty straightforward and I also have it auto-starting now on login. It has a server for network control, to which an iOS app called WiThrottle (free version) can connect. Installed that on an iPad and now can control the locomotive over wifi.

Not a bad outcome.

The next challenge is sound on board.

HO Trains and DCC++: Part 1

The Boy has rediscovered the model trains. Several years ago we started by resurrecting my 1990’s N-scale layout, which is 2’-x-4’ and could slide under the couch.  This was followed by  Dad’s post-war Lionel O-27, which was augmented by modern Fast Track and my 1970’s banjo crossing. The Fast Track was a game changer - the stuff works perfectly. It’s reliable, easy to join and doesn’t cause derailing. This was the first time the young-version Boy didn’t get frustrated while running trains. We added 21st century Thomas with a whistle. Finally we added an HO scale Thomas and Friends set. He was able to put together the track with this set himself. It also made several airplane trips in a carry-on back and forth to the Grandparents. But, then the XBox replaced the trains.

Most recently, the Boy pulled out the HO scale trains, built his own layout, and began running a Union Pacific (UP) diesel, boxcar and caboose. He’s using my 1980’s power supply, which was advanced at the time because it had a low-pass filter on the throttle to simulate gradual acceleration and deceleration of the train. Modern control, however, uses Digital Command and Control (DCC). The UP diesel has a DCC decoder. When we bought it, I figured I’d buy a command and base station soon after, but never found the opportune time.

Then last month the Boy has been talking about DCC nonstop and then he discovered locomotives with a Bluetooth interface. Time to do something about it. After researching DCC and reading the NMRA standard, I realized it could be implemented on an Arduino. Using my Google foo, lo and behold, I stumbled upon DCC++. I didn’t think it was the right time to invest in trains with dedicated Bluetooth interface and I have a couple Arduino Unos sitting around - this is the ticket.

Install DCC++

The DCC++ base station code can be downloaded from GitHub here. I used the standard Arduino IDE v1.6.5 to install it. The first time around compiling, it failed with the error:

Accessories.cpp:66:20: fatal error: EEPROM.h: 

No such file or directory #include <EEPROM.h>

Turns out in Arduino, all #includes no matter where they are used must be declared in the .ino file. I learned this interesting tidbit here. After adding #include <EEPROM.h> to the .ino file it works fine. I posted an issue to the GitHub site, so I imagine it'll be fixed. This is the first time I've contributed to someone else's code, although it is a pretty minor bug fix.

Prepare the Arduino Motor Shield

Per instructions here, I cut the indicated traces. The picture below shows the Vin trace cut - this disconnects the external power supply used to drive the motors through the motor shield from the Vin pin on the Arduino.

There are also a couple of jumpers to be installed. The final configuration is shown below.

Test the Uno

Before powering the motor shield, much less trying to control a train, I wanted to make sure I could talk to the DCC++ software on the Uno. Using Hyperterminal, I connected to the Arduino and sent the <s> status command. The response is shown in the picture.

Test the motor shield

Now that the software was working, I wanted to test the shield before using a train. I measured the voltage out when powered on. Using the <1> command turns the power on, which should be a 5-kHz square wave with almost 12-V amplitude (there might be some voltage drop in the driver chip). Note the LED’s by the terminal block are now also illuminated greenish-yellow. My digital volt meter is pretty decent at measuring root-mean-square (RMS) AC voltage and indicated 7.9 V for an expected 12/sqrt(2) = 8.5 V. (9/21/16: I was re-reading this and realized this statement is not correct. The RMS of a square wave is the amplitude. The sqrt(2) factor is for a sine wave. The low voltage reading on the voltmeter could be due it's frequency response. Or, it's possible there's a lot of loss in the driver. I need a scope to diagnose it.) That’ll do.

First run - throttle control

Now it was finally time to try it out. The DCC++ wiki gives an example command to move the train: <t 1 03 20 1>. This command tells locomotive “03,” which is the default encoder address, to move forward at speed 20. It worked. It worked without any fiddling. It’s nice when things just work. Though it is kind of anticlimactic. 

In Part 2, I’ll talk about our experience setting up iPad control.

ADC IIR LPF SPI TFT LCD GUI - Part 1

I have a project in which I need to read a signal, do some processing and display the results using a bar graph. After some fiddling with an Arduino UNO and researching the AVR microcontrollers they use, I settled on a two-MCU architecture. I use an Arduino Nano to acquire and process the data and an Arduino Mega to display the results and interface with the user. How many acronyms can be squeeze into a title?

ADC - analog-to-digital converter
IIR - infinite impulse response
LPF - low pass filter
SPI - serial peripheral interface
TFT - thin film transistor
LCD - liquid crystal display
GUI - graphical user interface

My first problem: how to get two Arduino’s to chat with each other. My work is based upon (copied from) a thorough and accurate forum post. The motivation to use SPI came from wanting to learn more about it. In researching the TVout library, I found someone generated an NTSC signal with the hardware based SPI master signal.  Given my current fascination with resurrecting my Atari days, I had an idea of trying to recreate a simplified ANTIC-like system on an Arduino in black and white. SPI seems the way to go there. But, back to my project.

The system here uses a Nano to read an analog voltage and apply a low pass filter. The Mega polls the Nano for the LPF output over SPI and displays the value on a bar graph. I offloaded the ADC work to the Nano thinking I would eventually use an interrupt-driven ADC to get a constant sampling rate. Because the TFT polls various analog channels to read the touchscreen, it seemed best to just off-load the other ADC work to a different microcontroller. The TFT fits on the Mega and leaves the header with the SPI interface unobstructed.

The SPI connections on the Nano are located on the ICSP header or pins D11-D13 and Slave Select is D10. The connections on the Mega are on the bottom header pins D50-D53. Connections are one-to-one. That is MISO:MISO, MOSI:MOSI, SCK:SCK, and SS:SS. The Master-In-Slave-Out MISO signal transfers data from the Nano to the Mega because the Nano is set up as SPI Slave and the Mega as Master. Vice-versa for MOSI. Serial Clock (SCK) sends a 2 MHz clock from the Mega to the Nano. And Slave Select (SS) signals the Nano to listen on SCK & MOSI for data and to send data on MISO synchronized to SCK. I use 2 Mbps because single-ended communications over long wires can’t go all that fast.

To low-pass filter the data, I use an exponential moving average filter implemented as an infinite impulse response (IIR) digital filter. This is very efficient because it requires only a weighted average of the current sample with the previous output. For a C++ implementation, I followed the integer implementation here. I had to add some additional 64-bit integer type casting to the coefficients in the filter equation to get it to operate correctly.. I also changed how the filter coefficient is implemented as a 16-bit unsigned integer (0-65535). The original author designed it to represent floats ranging from 1/65535 to 1 and I changed it to the range 0 to 65535/65536. Just personal preference. You can see the effect of the filter in the video - I adjust the potentiometer abruptly and it takes some time for the bargraph to totally respond.

Master Code

// Jeff Piepmeier - May 2016
//
// Main program adapted from SPI demo
// at http://www.gammon.com.au/spi 
// by Nick Gammon April 2011

//set up TFT display
#include <SPFD5408_Adafruit_GFX.h>
#include <SPFD5408_Adafruit_TFTLCD.h>
#include <SPI.h>

#define LCD_CS A3
#define LCD_CD A2
#define LCD_WR A1
#define LCD_RD A0
#define LCD_RESET A4

Adafruit_TFTLCD tft(LCD_CS, LCD_CD, LCD_WR, LCD_RD, LCD_RESET);

// variables for bar graph
int newHeight;
int oldHeight = 0;
int heightDiff;

void setup (void)
{
  // Serial.begin (115200);
  // Serial.println ("SPI demo");
  
  // SS = slave select, built in AVR output pin reference 
  digitalWrite(SS, HIGH);  // ensure SS stays high for now

  tft.reset();
  tft.begin(0x9341);
  tft.fillScreen(0x0000); // make the screen black
    
  // Put SCK, MOSI, SS pins into output mode
  // also put SCK, MOSI into LOW state, and SS into HIGH state.
  // Then put SPI hardware into Master mode and turn SPI on
  SPI.begin ();

  // Slow down the master a bit
  //SPI.setClockDivider(SPI_CLOCK_DIV4);
  // use 2 Mbps decided by testing. 4 Mbps has too many bit errors over 
  // jumper wires. Single-ended signals are not well suited for high-speed over wires
  SPI.beginTransaction(SPISettings(2000000, MSBFIRST, SPI_MODE0));

}  // end of setup

void loop (void)
{
  byte a=0; // variable to store SPI input data from Nano
  
  // get a value from the SPI - we're the master so have to ask for it from slave
  // enable Slave Select
  digitalWrite(SS, LOW);    
  a = SPI.transfer ('a');
  // disable Slave Select
  digitalWrite(SS, HIGH);

  newHeight=(int)(((long)a*319)/256); // hardcode display height of 320 lines
  heightDiff = oldHeight-newHeight; // only draw new part of bar graph for faster display
  if (heightDiff>0) { tft.fillRect(80, newHeight+1, 80, heightDiff+1, 0x0000); }
  else if (heightDiff<0) { tft.fillRect(80, oldHeight-1, 80, -heightDiff+1, 0xFFFF); }
  oldHeight=newHeight; // remember how high bar is
  
  // Serial.println (a, DEC);
}  // end of loop

Slave Code

// Main program is adapted from SPI demo at

// http://www.gammon.com.au/spi

// Written by Nick Gammon

// April 2011

//

// IIR Exponential Moving Average (EMA) Low Pass Filter (LPF)

// adapted from C++ code at

// http://stratifylabs.co/embedded%20design%20tips/2013/10/04/Tips-An-Easy-to-Use-Digital-Filter/

// filter coefficient float-to-uint16 conversion - min 0, max x=1 means 65535/65536=.9999847

#define DSP_EMA_I32_ALPHA(x) ( (uint16_t)(x * 65535) )

volatile byte command = 0;

volatile byte out1 = 0;

void setup (void)

{

  // have to send on master in, *slave out*

  pinMode(MISO, OUTPUT);

  digitalWrite(MISO, LOW); //ensure is low to start

  

  // turn on SPI in slave mode

  SPCR |= _BV(SPE);

  // turn on interrupts

  SPCR |= _BV(SPIE);

}  // end of setup

//http://stratifylabs.co/embedded%20design%20tips/2013/10/04/Tips-An-Easy-to-Use-Digital-Filter/

int32_t dsp_ema_i32(int32_t in, int32_t average, uint16_t alpha){

  int64_t tmp0;

  tmp0 = (int64_t)in * (int64_t)(alpha) + (int64_t)average * (int64_t)(65536 - alpha);

  return (int32_t)((tmp0 + 32768) / 65536);

}

// SPI interrupt routine

ISR (SPI_STC_vect)

{

  command = SPDR; // not yet used here

  SPDR = out1; 

}  // end of interrupt service routine (ISR) SPI_STC_vect

void loop (void)

{

  uint16_t adcReading = 0;

  static int32_t avg1 = 0;

  adcReading=analogRead(0); //10 bit unsigned, shift up to 31 bits for signed long int

  avg1=dsp_ema_i32( (int32_t)adcReading << 21, avg1, DSP_EMA_I32_ALPHA(0.0005));

  out1=byte(avg1 >> 23 ); // shift down to single byte

}  // end of loop

Easter 3D Print (and Splitting a 3D Model in Half)

Holidays are a good excuse to find something whimsical to 3D print. I found this pair of egg legs over on Thingiverse and chuckled. The legs are pose-able and hold an egg. Wife sketched an Humpty Dumpty on an egg and suggested I shoot the model in the garden. He can stand (above) or sit (below).

When I loaded up the model, I discovered the pieces would be difficult to print. Turns out the original designer is a grad student at MIT Media Lab and has access to a fancy industrial grade fabricator. On hobby printers, 3D objects are easiest to print when they have one flat face to serve as a base. For example, pyramids are easy to print and good for testing a printer setup. Otherwise, when an object has a part that hangs in mid air, the printing software can insert extra material for support. Sometimes though, some objects are just unprintable.

This model has 5 pieces: feet, lower legs, left & right upper legs, and a torso. The feet were easy. The lower legs printed OK with support. I rotated the upper legs 90 degrees and printed them with support. The "egg holding torso" print, however, failed miserably. To solve the problem, I was able to make my own flat face on the object by slicing it in half, printing the two halves, and super-gluing them together.

Surprisingly, there are few solutions for cutting 3D model objects in half to be googled-up. I ended up using Meshmixer, with which I've had varying success in the past. This time, however, it worked perfectly. Import the STL file and select "plane cut" under the "edit" menu. The default plane was oriented exactly along the major axes, splitting the part in half easily.

Print two, glue, build, pose, photograph, write a blog post. Happy Easter.

GRAVITEN (An Atari Gravitar rip-off in 10 lines of BASIC)

UPDATE April 3, 2016: 2nd Place!

GRAVITEN is my second entry to the 2016 NOMAM 10-liner BASIC game competition. Like NUTS!, this one is in the PUR-120 category. It's inspired by Gravitar's Red Planet 3 stage by Atari. I got the idea last Friday night playing Gravitar on my Atari 2600 10-in-1 joystick game. I had tried playing it as a kid, but found it too hard. Download my version here and run it on the Altirra emulator (800XL). (Direct link to ATR file.) If you play, please post your high-score in the comments!

Pilot your ship around the spiral to the planet's core. Reach the core and advance to the next level. Complete each level quickly to maximize points. But, be careful! Touch the wall and you lose a life.You get 3 lives to start and 1 additional life per level. Even numbered levels have atmospheric drag (indicated by gray background in the text window) and odd are in vacuum (indicated by black background). Gravity increases every two levels, starting at zero. Play through all 8 levels to complete the mission.  Controls: up to thrust, back to flip, left/right to rotate.

I learned a couple new things to write this game. I used Altirra's debug features to borrow the ship bitmap from Asteroids. Altirra can report the ANTIC register values allowing one to find the player-missile pixel graphics in memory. I also learned to use the TurboBasic XL parser tool (tbxl), which is really useful for at least two things: (1) storing binary data in strings using ATASCII characters and (2) optimizing code to squeeze into minimal space.

The Code

0DIMA$(96):A$=" ......................... {a whole bunch of ATASCII} ......................... ":A=.39269908:DIMA(1,15)
1F.B=0TO15:A(0,B)=SIN(A*B):A(1,B)=COS(A*B):N.B:GR.8:GR.5:POKE752,1:SE.0,3,4:SE.2,0,4:C.1:PAI.0,0:C.0:C=580:?"GRAVITEN"
2?"Stand by ";:A=.0174532925:F.B=0TOC STEP9:D=38*B/C*COS(B*A)+45:E=25*B/C*SIN(B*A)+15:F.F=1TO5:CI.D,E,F:N.F:?".";:N.B
3CLS:C.1:PL.0,0:DR.79,0:PL.19,39:DR.79,39:C.2:CI.46,15,1:G=44032:M.ADR(A$),G,80:POKE704,15:POKE54279,G/256:POKE53277,3
4POKE559,46:DO:?:?,"press FIRE to start";:W.STRIG(0):WE.:CLS:H=1e3:I=0:J=3:K=70:L=18:M=8:N=0:O=0:?:?I,J;:W.J:POKE657,22
5?H;" ";:M.G+512,G+513,127:M.M*5+G,G+512+INT(L),5:POKE53248,INT(K):P=Q:Q=STICK(0):R=(Q&4=4)-(Q&8=8):S=(P&2=2)&(Q&2=0)
6T=Q&1=0:M=(M+16+R+8*S)MOD16:IFT:SO.0,250,10,10:N=A(0,M)*.05+N:O=A(1,M)*.05+O:END.:IFI MOD2=0:N=N-.01*N:O=O-.01*O:END.
7N=I DIV2*2e-5*(135-K)+N:O=O-I DIV2*2e-5*(44-L):K=K+N:L=L-O:POKE53278,1:PA.1:U=PEEK(53252):IFU:K=70:L=18:M=8:N=0:O=0
8SO.0,50,U*8-6,15:PA.9:CLS:IFU=1:J=J-1:IFJ=0:?,"TRY AGAIN":END.:END.:IFU=2:H=(I+1)*1e3+H:IFI=7:?,"MISSION COMPLETE!":EX.
9END.:I=I+1:J=J+1:SE.0,I*2+3,4:SE.2,0,4-I MOD2*4:END.:?:?I,J,H;:END.:SO.0,0,0,0:H=H-1:WE.:?,"SCORE:";H:SO.0,0,0,0:LOOP

There are a couple of things worth pointing out in this code: the spiral path generation (in line 2) and the ship dynamics (in lines 6 & 7). The path is carved out of a solid field of color - this allows me to use missile-playfield collision detection to find out when the ship hits the wall. A TurboBASIC XL PAINT command is used to fill the screen. The spiral path is generated using a pair of parametric equations describing the polar coordinate equation r = aΘ. For each point on the curve, five concentric circles are drawn in the background color, which erase the foreground to create the path. A couple of lines are drawn to repair the top and bottom and to enclose the whole payfield in COLOR 1. Finally, a small box is plotted with the CIRCLE command in COLOR 2 for the goal.

The ship dynamics are borrowed from the earliest video games (Spacewar! and much later Asteroids), which based their game play on similar physics. Those two and other games used thrusters, gravity and drag in their game play. In Gravitar, there's gravity located in different places depending upon the scene. Here, the gravity is centered on the core (goal) and approximated so that it weakens as one progresses into the planet. The strength of the gravity is scaled by the level so higher levels become increasingly more difficult. I made it fairly weak because I'm not a very good gamer. In Asteroids, the ship encounters atmospheric drag - maybe it's the dust and debris ablated from the main rocks. I enable drag every other level. It's slightly easier to play with drag - it acts as a damper to over zealous thrust. You can almost point the ship where you want to go and hit the thursters. Without drag, once the ship starts moving in a direction, it doesn't change until a counter acting thrust is applied or gravity is present.

Thanks for reading.

NUTS! - Atari BASIC 10-Liner Contest 2016 Entry

Update April 3, 2016: 3rd Place!

Climb the trees, jump to gather acorns, but beware the blue jays! Earn points by climbing (press fire) and jumping left and right (move the joystick) to collect acorns, which only fall when you are climbing. The acorns are worth more the higher you climb, but there are more blue jays, too. The game is over when you run into a blue jay. Your high-score is recorded so you can try and beat it the next round. Grab the ATR here and play it on your favorite emulator (I use Altirra).

This is only my second BASIC program in modern times and is my entry to the NOMAM 2016 BASIC 10-Liner Contest. I like the 10-line constraint and find it a fun (for now) challenge to do my own code optimization squeezing it into 10 x120-character lines. For this program, I wrote it up in stages, getting each feature to work. I saved doing the pixel art for near the end. In the middle of development, I tested on the boy. He made me remove a timer:

"Dad, gamers hate timers in a runner, which is what this is." 

So, I removed the countdown timer and added an algorithm to increase difficulty with advancement. (More birds the more you climb.) After I compressed it down with single-character variable names, abbreviated statements, and some hand-optimized coding, I had space left over. I added sound effects in the remaining character count. I was surprised by how much more enjoyable the game play was with sound effects. Read on for the code. Here's a video of the final product. 

Here's the code listing in 10 lines each 120 characters or fewer:

0 DIMS$(76):S$="{...a mess of ATASCII...}":Q=ADR(S$):R=PEEK(106)
1 POKE106,R-8:GR.1:POKE106,R:CLS:POKE54279,R-4:POKE559,46:DP.53256,257:DP.53258,257:POKE53277,3:DP.708,$12C4:P=(R-4)*256
2 W=53248:DP.W,$6868:DP.W+2,$7888:DP.704,$850A:DP.706,$1D85:M.Q+64,DPEEK(560)+7,12:M.57344,(R-8)*256,1024:POKE756,R-8
3 M.Q+56,(R-8)*256+264,8:?#6;"NUTS!","HI:";J:?#6;"SCORE:":F.X=0TO10:?#6;" aaaaaa aaaaaa":N.X:F=1:S=16:H=104:B=1:E=0
4 D=12:G=16:K=0:Z=1:DO:POKE53278,1:A=2*STRIG(0):IFF=0:F=PTRIG(0)-PTRIG(1):END.:IFA=0:-M.P+664,P+666,78:-M.P+792,P+794,78
5 SO.1,0,Z,2:Z=Z=0:K=K+1:S=S-4:IFS<0:S=15:END.:IFB:D=D-2:M.Q+22-B*6+D,P+728-64*B,2:B=B*(D>0):EL.:B=(RND>(1-L))-(RND<L)
6 D=12:L=K/5E3:END.:END.:IFE:G=G-4+A:M.Q+40+G,P+920,4-A:E=G>0:EL.:G=16:E=(2-A)*(RND>.9):END.:-M.P+920,P+924-A,78:IFF=-1
7 IFH=104:F=0:EL.:H=H-8:M.Q,P+589,8:END.:END.:IFF=1:IFH=136:F=0:EL.:H=H+8:M.Q+8,P+589,8:END.:END.:POKEW,H:SO.0,0,0,0
8 X=PEEK(53260):SO.1,0,0,0:IFX&8:SO.0,50,10,15:K=K+K DIV 5:-M.P,P+964,28:ELSE:IFX&6:EXIT:END.:END.:POS.6,1:?#6;K;
9 SO.2,H,10+(F=0),6:PAUSE 0:POKE54277,S:LOOP:IFK>J:J=K:END.:SO.2,0,0,0:G.1

Below is the expanded code with commentary following each group of statements.

DIM S$(76)
S$="...{a bunch of ATASCII}..."
Q=ADR(S$)
Sets up a string full of ATASCII characters containing the pixel graphics for the acorn, squirrel, bluejay, and tree bark. The end contains characters making up part of the display list to enable vertical scrolling. The address of the string is stored in Q, which I use many times copy parts of the string into memory locations.

R=PEEK(106)
1 POKE 106,R-8
GRAPHICS 1
POKE 106,R
CLS 
Sets up the screen and creates blank memory for storing sprites and custom character set. The line #1 is used a jump at the end a game to restart another round.

POKE 54279,R-4: POKE 559,46: DPOKE 53256,257: DPOKE 53258,257: POKE 53277,3: DPOKE 708,$12C4: P=(R-4)*256: W=53248: DPOKE W,$6868: DPOKE W+2,$7888: DPOKE 704,$850A: DPOKE 706,$1D85
Sets up player-missile (sprite) graphics and colors. Turbo BASIC XL's double poke (abbreviated DP.) is great for saving code space when you need to set two adjacent 1-byte registers.

MOVE Q+64,DPEEK(560)+7,12
Modify the display list to make all but the first two rows Graphics 2 mode with vertical scrolling.

MOVE 57344,(R-8)*256,1024: POKE 756,R-8: MOVE Q+56,(R-8)*256+264,8
Copy the default character set into RAM and put a custom character into the "A" location. Point the hardware here.

? #6;"NUTS!","HI:";J: ? #6;"SCORE:": FOR X=0 TO 10: ? #6;" aaaaaa      aaaaaa": NEXT X
Draw the playfield onto the screen. The uppercase letters are printed in green. The lowercase "a" points to my special tree bark character with the brown color.

F=1: S=16: H=104: B=1: E=0: D=12: G=16: K=0: Z=1
Initialize a bunch of state variables, e.g. H is the horizontal position of the squirrel.

DO
Begin the main loop!

POKE 53278,1
Clear the collision register.

A=2*STRIG(0)
IF F=0:F=PTRIG(0)-PTRIG(1):ENDIF
Get joystick input. Only read the left/right direction if the squirrel is not jumping (F=0).

IF A=0
Now begins a large set of actions if the user has the fire button pressed:

-MOVE P+664,P+666,78: -MOVE P+792,P+794,78
These two move commands scroll the blue jays down the screen.

SOUND 1,0,Z,2: Z=Z=0
Play the climbing sound and toggle it on and off with flag Z. This takes advantage of the odd numbered distortion values being no sound.

K=K+1
Increase the score by one for climbing.

S=S-4: IF S<0: S=15: ENDIF
Smooth scroll 1/4 of a character to make it look like the squirrel is going up the tree.

IF B: D=D-2: MOVE Q+22-B*6+D,P+728-64*B,2: B=B*(D>0)
If there's a bird being scrolled onto the tree ... then put two lines of the sprite onto the screen at a time. Turn off B when the whole bird makes it.

ELSE: B=(RND>(1-L))-(RND<L): D=12: L=K/5000
Otherwise test to see if there's a new bird. Compute the likelihood based on the score.

ENDIF: ENDIF
End of the bird conditional (IF B). End of the climbing conditional (IF A=0).

IF E: G=G-4+A: MOVE Q+40+G,P+920,4-A: E=G>0
If there's an acorn entering the screen ... introduce either 2 or 4 lines of the acorn at a time depending upon the climbing state.

ELSE: G=16: E=(2-A)*(RND>0.9)
Otherwise figure out if we need another acorn, but only if the squirrel is climbing. This prevents a player from parking out and collecting acorns with no other challenge.

ENDIF : -MOVE P+920,P+924-A,78
Scroll the acorns down the screen.

IF F=-1
  IF H=104
    F=0
  ELSE
    H=H-8: MOVE Q,P+589,8
  ENDIF
ENDIF
IF F=1
  IF H=136
    F=0
  ELSE
    H=H+8: MOVE Q+8,P+589,8
  ENDIF
ENDIF
POKE W,H
This section jumps the squirrel left and right. It moves the player in 8 columns increments and selects which pixel graphic to display (either left or right facing squirrel). This was some of the first loop code I wrote. There may be a more compact way to do this with math and logical expressions, but these two IF-THEN structures do the trick and don't take up too much space. I suspect this is a bit faster because there's no multiplications that would be required in a more compact approach.

SOUND 0,0,0,0: SOUND 1,0,0,0
Part of the sound effect logic - here the jumping and climbing sounds are turned off. 

X=PEEK(53260)
IF X&8
  SOUND 0,50,10,15: K=K+K DIV 5: -MOVE P,P+964,28
ELSE :IF X&6
  EXIT :ENDIF
ENDIF
Check the collision register for either an acorn or blue jay hit. If an acorn, the play a tone, increment the score (by 20%), and erase the acorn. If a blue jay, exit the DO loop.

POSITION 6,1: ? #6;K
Update the score.

SOUND 2,H,10+(F=0),6
Play a sound with pitch based on the horizontal position of the squirrel when it is jumping. These sound effects were squeezed in at the end, which accounts for the inconsistent way they are implemented.

REPEAT :UNTIL PEEK(54283)>93
PAUSE 0
POKE 54277,S
Poor man's vertical blank interrupt (VBI). Since the rules prohibit machine code, just hand around until the VCOUNT register is mostly down the screen and then update the fine scrolling register. 
Call a PAUSE routine to sync up the code with the vertical blank (PAUSE counts v-blanks to keep time). This prevents flicker on the bottom row and tearing of the squirrel sprite while jumping. 

LOOP
End of the game loop - go back and DO it all again.

IF K>J: J=K: ENDIF
When a squirrel hits a blue jay, the EXIT shifts execution to here. Update the high score.

SOUND 2,0,0,0
Turn off the jumping sound. The other sounds are already turned off.

GOTO 1
Restart the game without resetting the high score.